WO2021258644A1 - Indoor environment health degree regulating method and system based on machine vision - Google Patents

Abstract

An indoor environment health degree regulating method and system based on machine vision. The method comprises the following steps: (1) collecting facial data of a person, and applying independent vector analysis to analyze a periodic signal from the facial data to detect the heart rate; (2) preprocessing heart rate data collected in a health environment to construct an environment health degree language word model; and (3) comparing heart rate data collected in an actual environment with data in the constructed environment health degree language word model to determine whether the environment is healthy or not.

Description

Indoor environment health adjustment method and system based on machine vision Technical field

The invention relates to a method for judging the health of an indoor environment, in particular to an adjusting method for judging the health of an indoor environment based on machine vision. The method relates to the field of smart home technology.

Background technique

The statements here only provide background art related to the present invention, and do not necessarily constitute prior art.

In recent years, people have paid more and more attention to the comfort and health of the living environment. According to survey statistics, people spend more than 80% of their time indoors every day, and indoor temperature has an important impact on people's health. Therefore, a suitable indoor environment temperature is an important factor in judging the health of the indoor environment.

At present, when the indoor environment temperature is detected, sensor detection is often used, and then the collected value is compared with a subjectively set value to determine whether the environment is healthy. Due to the difference of each person's physique and age, the need for environmental temperature is also different. Therefore, subjectively set values cannot accurately determine whether the environment is healthy. Secondly, the existing heart rate detection methods need to be in contact with the human body or wear a device. It can be seen from these problems that the prior art lacks a more convenient and accurate method for judging the health of the indoor living environment.

Summary of the invention

In order to judge the health of the indoor living environment more conveniently and accurately, the present invention proposes a method and system for adjusting the health of the indoor environment based on machine vision.

In order to achieve the above objectives, the present invention is achieved through the following technical solutions:

The present invention provides a method for adjusting indoor environment health based on machine vision, which includes the following steps:

(1) Collect human face data, apply independent vector analysis to analyze the periodic signal from the face data, so as to detect the heart rate;

(2) Preprocess the collected heart rate data in a healthy environment to construct an environmental health language word model;

(3) Collect the heart rate data in the actual environment and compare it with the data in the constructed environmental health model to determine whether the environment is healthy.

Preferably, the steps of the step (1) are as follows:

With the help of a high-definition camera to take real-time data of multiple skin areas on the human face, the remote photoplethysmography heart rate monitoring method based on the joint blind source separation algorithm is adopted, and the independent vector is used for joint analysis to obtain the human heart rate data.

Preferably, the specific steps of the step (1) are as follows:

First, select the skin area for data collection; then calculate the spatial mean of the color RGB from the collected skin data; second, apply the signal processing method to the calculated spatial mean to obtain the components of each skin area containing the heart rate information; third, Use independent vector analysis to extract the common signal components of different mixed signal groups; finally, fast Fourier transform is applied to this component in order to estimate the corresponding frequency or the number of peaks Ns during the processing duration T(s). The heart rate in beats per minute will be calculated as 60×Fs or Ns/T×60.

Preferably, the specific steps of the step (2) are as follows:

Step 1: Convert monitoring data to interval data

1) Statistical calculation of daily acquired data:

Assuming that the data collected on the i day is processed, the sample mean _mi and sample standard deviation σ _i are first calculated, which are expressed as:

Where n _i represents the total amount of data collected on the i day, and data _{i, j} represents the jth data collected on the i day;

2) Daily data preprocessing:

On the basis of stage 1) _{, judge whether each data i, j} satisfies the following equation:

|data _i,j -m _i |≤k*σ _i (3)

If the equation is satisfied, then accept; otherwise, it will be eliminated; k represents the constraint coefficient, generally k is 2; after this processing, the data in the i-th day will be left n″ _i (n″ _{i ≤ n i} ) Piece;

3) Statistical calculation of all data left in n days:

Calculate the sample mean m and sample standard deviation σ of all remaining data in n days:

4) Preprocessing of data within n days: _{judge whether each data i, j} satisfies equation (3), and the data that meets the requirements are accepted, otherwise, it will be rejected;

5) Get the daily interval:

In the data collected every day, the maximum and minimum values are selected to form a daily interval, and the interval on the i-th day is expressed as:

Where i = 1, ..., n, it represents the number of days after the data preprocessing stage above left, c _i, and D _i respectively represent the left end of the i-th day intervals day and right points;

Step 2: Interval data preprocessing

1) abnormal value processing: c _i is first performed on the d _i Box and Whisker test and then calculate L _i = c _i -d _i; if the inclusive interval satisfies the following equation:

The interval is retained, otherwise it is eliminated; Q (.25) is called the lower four-digit score, which means that a quarter of the data value of all observations is smaller than it; Q (.75) is called the upper four-digit score , Which means that a quarter of the data values in all observations are larger than it; IQR is called the interquartile range, which is the difference between the upper four scores and the lower four scores;

After this treatment, there m'≤n data interval is retained; calculate c _i, d _i, and L _i of the sample mean and standard deviation, as _{(m c, σ c),} (m d, σ d), (m _L ,σ _L ), where i=1,...,m′;

2) Tolerance value processing: If the endpoint values of the retained m'data intervals satisfy the following equation:

The interval is reserved; otherwise, it will be kicked out. Where i=1,...,m′, k represents the constraint coefficient, and the value of k is 2;

Thereafter, m "≤n data interval is retained; recalculate the sample mean and standard deviation c _i, d _i and L _i, such as _{(m c ', σ c'} ), (m d ', σ d') ,(m _L ′,σ _L ′), where i=1,...,m″;

3) Reasonable treatment: calculation

ξ*={(m _c ′(σ′ _d ) ² -m _d ′(σ′ _c ) ² )±σ _c ′σ _d ′[(m _c ′-m _d ′) ² +2((σ′ _c ) ² -(σ′ _d ) ² )ln(σ _c ′/σ _d ′)] ^1/2 }/((σ′ _c ) ² -(σ′ _d ) ² ) (9)

When m _c ′≤ξ*≤m _d ′, the interval will be retained; otherwise, the interval will be eliminated;

Renumber the reserved n′(1≤n′≤n) data intervals as 1, 2,...,n′, and express them as [t _i ^l ,t _i ^r ],(i=1,2, ...,n′);

Step 3: Build a language word model for environmental health

Apply the percentile method to the retained n′(1≤n′≤n) intervals to select two representative intervals, and construct an environmental health language word model;

Assume that the left and right endpoints of the retained interval data are arranged as

For a given q (q<0.5), suppose the 100th and 100th (1-q) percentiles are respectively represented as [T _q ,T _1-q ], this interval contains the (1-2q) ratio Data points. For the left endpoint, the 100th and 100th (1-q) percentiles are calculated as

T _q ^L = t ^l _[n′*q] +rem(n′*q,1)(t ^l _[n′*q+1] -t ^l _[n′*q] ) (10)

in

and

Respectively represent the 100th and 100th (1-q) percentiles of the left end point, [.] uses the floor function to represent the integral part of the corresponding value, and rem(·, 1) uses the mod function to calculate the corresponding value after dividing by 1 remainder. Similarly, for the right endpoint, the 100th and 100th (1-q) percentiles can be calculated and expressed as

and

T _q ^R = t ^r _[n′*q] +rem(n′*q,1)(t ^r _[n′*q+1] -t ^r _[n′*q] ) (12)

Suppose the left and right representative intervals of the environmental health language word model are

Construct a language word model for environmental health.

Preferably, in the step (3), the facial data in the actual environment is collected, and then the heart rate recognition module is called, and the super-sensing heart rate monitoring method based on the joint blind source separation algorithm is applied to jointly analyze the facial data to obtain the Heart rate data, and then compare the data with the environmental health language word model to determine whether the environmental temperature is high or low, so that the air conditioner can act accordingly.

Preferably, the specific judgment rules in step (3) are as follows:

Suppose the monitored heart rate value of the actual environment is x;

(1) x＜＜LL, the ambient temperature is very low, and the air conditioner setting temperature rises to the RR value;

(2) x<LL, the ambient temperature is low, and the air conditioner set temperature increases by the value of RL;

(3) LL＜x＜LR, the ambient temperature is more comfortable, and the air conditioner temperature rises by 1 degree;

(4) LR≤x≤RL, the ambient temperature is comfortable, and the air-conditioning set temperature remains unchanged;

(5) RL＜x＜RR, the ambient temperature is more comfortable, and the air-conditioning set temperature drops by 1 degree;

(6) RR<x, the ambient temperature is higher, and the air-conditioning set temperature drops to the LR value;

(7) RR<<x, the ambient temperature is very high, and the air conditioner set temperature drops to the LL value.

The present invention also provides an indoor environment health adjustment system based on machine vision, which is used to execute the steps of the above-mentioned machine vision-based indoor environment health adjustment method, including:

A heart rate recognition module, which is used to perform the method of step (1);

Environmental health modeling module, which is used to execute the method of step (2);

Environmental health discrimination and adjustment module, which is used to execute the method of step (3).

The technical scheme of the present invention has the following beneficial effects:

1. The heart rate monitoring adopts the super-sensing method, and the data collection is faster and more convenient, which improves the intelligence of the home.

2. The remote photoplethysmography heart rate monitoring method based on the combined blind source separation algorithm analyzes multiple sub-regions of the face, which can overcome the influence of light changes and exercise and improve the accuracy of the heart rate monitoring value.

3. Quantify people's "feeling" through the heart rate, eliminating the interference of subjective consciousness, the heart rate is relatively stable, and the judgment result is more accurate.

Description of the drawings

The accompanying drawings of the specification constituting a part of the present invention are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and the description thereof are used to explain the present invention, and do not constitute an improper limitation of the present invention.

Figure 1 is a diagram of the environmental health language word model of the present invention;

Fig. 2 is a flow chart of judging and adjusting the indoor environment health of the present invention.

detailed description

It should be pointed out that the following detailed descriptions are all illustrative and are intended to provide further descriptions of the present invention. Unless otherwise specified, all technical and scientific terms used in the present invention have the same meanings as commonly understood by those of ordinary skill in the technical field to which the present invention belongs.

It should be noted that the terms used here are only for describing specific embodiments, and are not intended to limit the exemplary embodiments according to the present invention. As used herein, unless the present invention clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should also be understood that when the terms "comprising" and/or "including" are used in this specification, they Indicate the existence of features, steps, operations, devices, components, and/or their combinations;

In order to judge the health of the indoor living environment more conveniently and accurately, a method and system for adjusting the health of the indoor environment based on machine vision is proposed. The high-definition camera is used to collect facial data of people living in a healthy environment for a period of time, and the heart rate data is analyzed from the facial data using the remote photoplethysmography heart rate monitoring method based on the joint blind source separation algorithm, and then the data is analyzed Preprocessing, and then convert the collected data into interval data, and then process the interval data, and build an environmental health language word model on this basis. Then, collect the human heart rate data in the actual environment and compare the data with the data in the environmental health language word model to determine whether the environmental temperature is high or low, and then give an adjustment strategy.

The invention is composed of a heart rate recognition module, an environmental health degree modeling module, and an environmental health degree discrimination and adjustment module. The heart rate recognition module mainly uses a high-definition camera to collect human facial data, and uses independent vector analysis to analyze periodic signals from the facial data to detect the heart rate. The environmental health modeling module is to call the heart rate recognition module to identify the heart rate in a healthy environment, and then preprocess the heart rate data to construct an environmental health language word model. The environmental health discrimination and adjustment module collects facial data in the actual environment, calls the heart rate recognition module to obtain the heart rate data, compares it with the data in the constructed environmental health model, determines whether the environment is healthy, and gives appropriate adjustment strategies .

1. Heart rate recognition module

The function of this module is to analyze the person's heart rate data from the person's facial data. With the help of a high-definition camera to take real-time data of multiple skin areas on the human face, the remote photoplethysmography heart rate monitoring method based on the joint blind source separation algorithm is adopted, and the independent vector is used for joint analysis to obtain the human heart rate data.

First, select the skin area for data collection; then calculate the spatial mean of the color RGB from the collected skin data; second, apply the signal processing method to the calculated spatial mean to obtain the components of each skin area containing the heart rate information; third, Use independent vector analysis to extract the common signal components of different mixed signal groups; finally, fast Fourier transform is applied to this component in order to estimate the corresponding frequency (or the number of peaks Ns during the processing duration T(s)). The heart rate in beats per minute will be calculated as 60×Fs (or Ns/T×60).

2. Environmental Health Modeling Module

Use a high-definition camera to collect facial data of a person in a healthy environment, call the heart rate recognition module to obtain the heart rate data at the corresponding time, and then preprocess the collected heart rate data. Then the daily heart rate data is converted into a heart rate interval, and then the interval data is preprocessed in three steps: singular value, tolerance limit, and reasonable value. The processed interval data is constructed into a language word model of environmental health by using the percentile method. details as follows:

1. Convert monitoring data to interval data

(1) Statistical calculation of daily acquired data:

Assuming that the data collected on the i day is processed, the sample mean _mi and sample standard deviation σ _i are first calculated, which are expressed as:

Where n _i represents the total amount of data collected on the i day, and data _{i, j} represents the jth data collected on the i day;

(2) Daily data preprocessing:

On the basis of stage (1) _{, judge whether each data i, j} satisfies the following equation:

|data _i,j -m _i |≤k*σ _i (3)

If the equation is satisfied, then accept; otherwise, it will be eliminated; k is the constraint coefficient, generally k is 2; after this processing, the data in the i-th day will be left n″ _i (n″ _{i ≤ n i} ) Piece;

(3) Statistical calculation of all data left in n days:

Calculate the sample mean m and sample standard deviation σ of all remaining data in n days:

(4) Data preprocessing within n days: _{judge whether each data i, j} satisfies equation (3), and the data is accepted if it satisfies it, otherwise, it is rejected;

(5) Get the daily interval:

In the data collected every day, the maximum and minimum values are selected to form a daily interval, and the interval on the i-th day is expressed as:

Where i = 1, ..., n, it represents the number of days after the data preprocessing stage above left, c _i, and D _i respectively represent the left end of the i-th day intervals day and right points;

2. Interval data preprocessing

(1) handling an abnormal value: c _i is first performed on the d _i Box and Whisker test and then calculate L _i = c _i -d _i; if the inclusive interval satisfies the following equation:

The interval is retained, otherwise it is eliminated; Q (.25) is called the lower four-digit score, which means that a quarter of the data value of all observations is smaller than it; Q (.75) is called the upper four-digit score , Which means that a quarter of the data values in all observations are larger than it; IQR is called the interquartile range, which is the difference between the upper four scores and the lower four scores;

After this treatment, there m'≤n data interval is retained; calculate c _i, d _i, and L _i of the sample mean and standard deviation, as _{(m c, σ c),} (m d, σ d), (m _L ,σ _L ), where i=1,...,m′;

(2) Tolerance value processing: If the endpoint values of the retained m'data intervals satisfy the following equation:

Then the interval is reserved; otherwise, it will be kicked out; where i=1,...,m′, k represents the constraint coefficient, and the value of k is 2;

Thereafter, m "≤n data interval is retained; recalculate the sample mean and standard deviation c _i, d _i and L _i, such as _{(m c ', σ c'} ), (m d ', σ d') ,(m _L ′,σ _L ′), where i=1,...,m″;

3) Reasonable treatment: calculation

ξ*={(m _c ′(σ′ _d ) ² -m _d ′(σ′ _c ) ² )±σ _c ′σ _d ′[(m _c ′-m _d ′) ² +2((σ′ _c ) ² -(σ′ _d ) ² )ln(σ _c ′/σ _d ′)] ^1/2 }/((σ′ _c ) ² -(σ′ _d ) ² ) (9)

When m _c ′≤ξ*≤m _d ′, the interval will be retained; otherwise, the interval will be eliminated;

Renumber the reserved n′(1≤n′≤n) data intervals as 1, 2,...,n′, and express them as [t _i ^l ,t _i ^r ],(i=1,2, ...,n′);

3. Building a language word model for environmental health

Apply the percentile method to the retained n′(1≤n′≤n) intervals to select two representative intervals, and construct an environmental health language word model;

Assume that the left and right endpoints of the retained interval data are arranged as

For a given q (q<0.5), suppose the 100th and 100th (1-q) percentiles are respectively represented as [T _q ,T _1-q ], this interval contains the (1-2q) ratio For the left endpoint, the 100th and 100th (1-q) percentiles are calculated as

T _q ^L = t ^l _[n′*q] +rem(n′*q,1)(t ^l _[n′*q+1] -t ^l _[n′*q] ) (10)

in

and

Respectively represent the 100th and 100th (1-q) percentiles of the left end point, [.] uses the floor function to represent the integral part of the corresponding value, and rem(·, 1) uses the mod function to calculate the corresponding value after dividing by 1 remainder. Similarly, for the right endpoint, the 100th and 100th (1-q) percentiles can be calculated and expressed as

and

T _q ^R = t ^r _[n′*q] +rem(n′*q,1)(t ^r _[n′*q+1] -t ^r _[n′*q] ) (12)

Suppose the left and right representative intervals of the environmental health language word model are

Then build the environmental health language word model as shown in Figure 1.

3. Environmental Health Judgment Module

Collect the facial data in the actual environment, then call the heart rate recognition module, apply the super-sensing heart rate monitoring method based on the joint blind source separation algorithm to jointly analyze the facial data to obtain the heart rate data in the environment, and then compare the data with the health of the environment The degree language word model (Figure 1) is compared to determine whether the ambient temperature is high or low, and the air conditioner can act accordingly. The specific judgment rules are as follows: set the monitored heart rate value of the actual environment as x;

(1) x＜＜LL, the ambient temperature is very low, and the air conditioner setting temperature rises to the RR value;

(2) x<LL, the ambient temperature is low, and the air conditioner set temperature increases by the value of RL;

(3) LL＜x＜LR, the ambient temperature is more comfortable, and the air conditioner temperature rises by 1 degree;

(4) LR≤x≤RL, the ambient temperature is comfortable, and the air-conditioning set temperature remains unchanged;

(5) RL＜x＜RR, the ambient temperature is more comfortable, and the air-conditioning set temperature drops by 1 degree;

(6) RR<x, the ambient temperature is higher, and the air-conditioning set temperature drops to the LR value;

(7) RR<<x, the ambient temperature is very high, and the air conditioner set temperature drops to the LL value.

The overall steps of the present invention are shown in Figure 2.

The above descriptions are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention can have various modifications and changes. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims (7)

1. A method for adjusting indoor environment health based on machine vision, which is characterized in that it comprises the following steps:

   (1) Collect human face data, apply independent vector analysis to analyze the periodic signal from the face data, so as to detect the heart rate;

   (2) Preprocess the collected heart rate data in a healthy environment to construct an environmental health language word model;

   (3) Collect the heart rate data in the actual environment and compare it with the data in the constructed environmental health model to determine whether the environment is healthy.
2. The method for adjusting indoor environment health based on machine vision according to claim 1, wherein the steps of step (1) are as follows:

   With the help of a high-definition camera to take real-time data of multiple skin areas on the human face, the remote photoplethysmography heart rate monitoring method based on the joint blind source separation algorithm is adopted, and the independent vector is used for joint analysis to obtain the human heart rate data.
3. The method for adjusting indoor environment health based on machine vision according to claim 2, wherein the specific steps of the step (1) are as follows:

   First, select the skin area for data collection; then calculate the spatial mean of the color RGB from the collected skin data; second, apply the signal processing method to the calculated spatial mean to obtain the components of each skin area containing the heart rate information; third, Use independent vector analysis to extract the common signal components of different mixed signal groups; finally, apply the fast Fourier transform to this component to estimate the corresponding frequency or the number of peaks Ns during the processing duration T(s); in the form of beats per minute The heart rate will be calculated as 60×Fs or Ns/T×60.
4. The indoor environment health adjustment method based on machine vision according to claim 1, wherein the specific steps of the step (2) are as follows:

   Step 1: Convert monitoring data to interval data

   1) Statistical calculation of daily acquired data:

   Assuming that the data collected on the i day is processed, the sample mean _mi and sample standard deviation σ _i are first calculated, which are expressed as:

   

   

   Where n _i represents the total amount of data collected on the i day, and data _{i, j} represents the jth data collected on the i day;

   2) Daily data preprocessing:

   On the basis of stage 1) _{, judge whether each data i, j} satisfies the following equation:

   |data _i,j -m _i |≤k*σ _i (3)

   If the equation is satisfied, then accept; otherwise, it will be eliminated; k represents the constraint coefficient, after this processing, the data of the i-th day will be left n″ _i (n″ _{i ≤ n i} );

   3) Statistical calculation of all data left in n days:

   Calculate the sample mean m and sample standard deviation σ of all remaining data in n days:

   

   

   4) Preprocessing of data within n days: _{judge whether each data i, j} satisfies equation (3), and the data that meets the requirements are accepted, otherwise, it will be rejected;

   5) Get the daily interval:

   In the data collected every day, the maximum and minimum values are selected to form a daily interval, and the interval on the i-th day is expressed as:

   

   Where i = 1, ..., n, it represents the number of days after the data preprocessing stage above the left; C _i and D _i respectively represent the left end of the i-th day intervals day and right points;

   Step 2: Interval data preprocessing

   1) abnormal value processing: c _i is first performed on the d _i Box and Whisker test and then calculate L _i = c _i -d _i; if the inclusive interval satisfies the following equation:

   c _i ∈[Q _c (.25)-1.5IQR _c ,Q _c (.75)+1.5IQR _c ]

   d _i ∈[Q _d (.25)-1.5IQR _d ,Q _d (.75)+1.5IQR _d ] (7)

   L _i ∈[Q _L (.25)-1.5IQR _L ,Q _L (.75)+1.5IQR _L ]

   The interval is retained, otherwise it is eliminated; Q (.25) is called the lower four-digit score, which means that a quarter of the data value of all observations is smaller than it; Q (.75) is called the upper four-digit score , Which means that a quarter of the data values in all observations are larger than it; IQR is called the interquartile range, which is the difference between the upper four scores and the lower four scores;

   After this treatment, there m'≤n data interval is retained; calculate c _i, d _i, and L _i of the sample mean and standard deviation, as _{(m c, σ c),} (m d, σ d), (m _L ,σ _L ), where i=1,...,m′;

   2) Tolerance value processing: If the endpoint values of the retained m'data intervals satisfy the following equation:

   c _i ∈[m _c -kσ _c ,m _c +kσ _c ]

   d _i ∈[m _d -kσ _d ,m _d +kσ _d ] (8)

   L _i ∈[m _L -kσ _L ,m _L +kσ _L ]

   Then the interval is reserved; otherwise, it will be kicked out; where i=1,...,m′, k represents the constraint coefficient, and the value of k is 2;

   Thereafter, m "≤n data interval is retained; recalculate the sample mean and standard deviation c _i, d _i and L _i, such as _{(m c ', σ c'} ), (m d ', σ d') ,(m _L ′,σ _L ′), where i=1,...,m″;

   3) Reasonable treatment: calculation

   ξ*={(m _c ′(σ′ _d ) ² -m _d ′(σ′ _c ) ² )±σ _c ′σ _d ′[(m _c ′-m _d ′) ² +2((σ′ _c ) ² -(σ′ _d ) ² )ln(σ _c ′/σ _d ′)] ^1/2 }/((σ′ _c ) ² -(σ′ _d ) ² ) (9)

   When m _c ′≤ξ*≤m _d ′, the interval will be retained; otherwise, the interval will be eliminated;

   Renumber the reserved n′(1≤n′≤n) data intervals as 1, 2,...,n′, and express them as [t _i ^l ,t _i ^r ],(i=1,2, ...,n′);

   Step 3: Build a language word model for environmental health

   Apply the percentile method to the retained n′(1≤n′≤n) intervals to select two representative intervals, and construct an environmental health language word model;

   Assume that the left and right endpoints of the retained interval data are arranged as

   

   

   For a given q, suppose the 100th and 100th (1-q) percentiles are respectively represented as [T _q ,T _1-q ], this interval contains the data points of the (1-2q) ratio; For the left endpoint, the 100th and 100th (1-q) percentiles are calculated as

   T _q ^L = t ^l _[n′*q] +rem(n′*q,1)(t ^l _[n′*q+1] -t ^l _[n′*q] ) (10)

   

   in

   

   and

   

   Respectively represent the 100th and 100th (1-q) percentiles of the left end point, [.] uses the floor function to represent the integral part of the corresponding value, and rem(·, 1) uses the mod function to calculate the corresponding value after dividing by 1 Remainder; Similarly, for the right endpoint, the 100th and 100th (1-q) percentiles can be calculated and expressed as

   

   and

   

   T _q ^R = t ^r _[n′*q] +rem(n′*q,1)(t ^r _[n′*q+1] -t ^r _[n′*q] ) (12)

   

   Suppose the left and right representative interval of the environmental health language word model is

   

   

   Construct a language word model for environmental health.
5. The indoor environment health adjustment method based on machine vision according to claim 1, characterized in that, in the step (3), facial data in the actual environment is collected, and then the heart rate recognition module is called, and the algorithm based on joint blind source separation is applied The super-sensing heart rate monitoring method jointly analyzes the face data to obtain the heart rate data in the environment, and then compares the data with the environmental health language word model to determine whether the environment temperature is high or low, so that the air conditioner Make corresponding actions.
6. The method for adjusting the health of an indoor environment based on machine vision according to claim 5, characterized in that the specific judgment rules of the step (3) are as follows:

   Suppose the monitored heart rate value of the actual environment is x

   (1) x＜＜LL, the ambient temperature is very low, and the air conditioner setting temperature rises to the RR value;

   (2) x<LL, the ambient temperature is low, and the air-conditioning set temperature rises by the value of RL;

   (3) LL<x<LR, the ambient temperature is more comfortable, and the air-conditioning temperature rises by 1 degree;

   (4) LR≤x≤RL, the ambient temperature is comfortable, and the air-conditioning set temperature remains unchanged;

   (5) RL<x<RR, the ambient temperature is more comfortable, and the air-conditioning set temperature drops by 1 degree;

   (6) RR<x, the ambient temperature is higher, and the air conditioner set temperature drops to the LR value;

   (7) RR<<x, the ambient temperature is very high, and the air conditioner set temperature drops to the LL value.
7. A machine vision-based indoor environment health adjustment system, which is characterized in that it is used to implement the steps of the machine vision-based indoor environment health adjustment method according to any one of claims 1-6 when executed, comprising:

   A heart rate recognition module, which is used to perform the method of step (1);

   Environmental health modeling module, which is used to execute the method of step (2);

   Environmental health discrimination and adjustment module, which is used to execute the method of step (3).

Images